一种融合遮挡分割的多目标跟踪算法

A Multi-target Tracking Algorithm Combined with Occlusion Segmentation

复杂环境下的多目标视频跟踪是计算机视觉领域的一个难点,有效处理目标间遮挡是解决多目标跟踪问题的关键。提出了一种融合遮挡分割的多目标跟踪算法,计算每个目标的光流速度概率直方图,反映其运动统计信息;综合使用外观、运动、颜色信息构造新的像素距离表达,借助分阶段分类思想及K均值聚类技术进行遮挡分割,得到准确的运动前景像素;在粒子滤波器跟踪框架下,使用概率外观模型进行多目标跟踪,更好地处理动态遮挡问题。实验表明,所提算法解决了复杂环境下的多目标跟踪问题。

Multi-target tracking in complex scenes is one of the most complicated problems in computer vision. Handling the occlusion between objects is the key issue in multi-target tracking. An occlusion segmentation-based algorithm is presented to track multiple people through complex situations which are captured by static monocular cameras. In the proposed algorithm, the probabilistic histogram of each object＇s optical flow vector is calculated, then this motion statistic information along with the color and appearance information is used to construct a new expression of pixel distance, finally stepwise classification and K-means clustering method are taken advantages of to accomplish occlusion segmentation. Object is handled by a particle filter-based tracking framework, and a probabilistic appearance model is used to find the best particle. Results show that the proposed approach can improve the performance of the original probabilistic appearance model and handle dynamic occlusion better.